Microbiological Safety of Industrially Reared Insects for Food: Identification

Home , Axenic, Endospore

bioRxiv preprint doi: https://doi.org/10.1101/2020.04.22.055236; this version posted April 23, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

1 Microbiological safety of industrially reared insects for food: Identification

2 of bacterial endospores and targeted detection of foodborne viruses

3 Vandeweyera, D., Lievensb, B., Van Campenhouta*, L.

4 a KU Leuven, Department of Microbial and Molecular Systems (M²S), Lab4Food, Geel

5 Campus, 2440 Geel, Belgium

6 b KU Leuven, Department of Microbial and Molecular Systems (M²S), Laboratory for Process

7 Microbial Ecology and Bioinspirational Management (PME&BIM), Group T Campus, 3000

8 Leuven, Belgium

23 * Corresponding author: Leen Van Campenhout, KU Leuven, Geel Campus, Kleinhoefstraat 4,

24 2440 Geel, Belgium. E-mail: [email protected] 1

25 Abstract

26 Edible insects are characterised by high microbial numbers of which the bacterial endospores

27 fraction can survive thermal processing. It is unknown, however, which bacterial species occur

28 as endospore in edible insects and what impact they have on food safety. Additionally, edible

29 insects have never been explored for the presence of foodborne viruses so far. In this study, we

30 found that the bacterial endospore fraction in a collection of mealworm and cricket samples

31 obtained from commercial insect producers can comprise a large amount of Bacillus cereus

32 group members that can pose insect or human health risks. Monitoring and effective mitigation

33 of these risks are urged. By contrast, none of the foodborne viruses hepatitis A virus, hepatitis

34 E virus and norovirus genogroup II were detected in the sample collection. Therefore, food

35 safety risks originating from these viral pathogens can be concluded to be low.

36 Introduction

37 As from 2014, edible insects and insect-derived food products started to enter the

38 European markets. According to the International Platform of Insects for Food and Feed

39 (IPIFF), 6,000 tonnes of insect proteins were produced by its European members in 2019. The

40 organisation estimates production to be around 3 mio tonnes in 2030. Due to the expansion of

41 the insect sector risk assessments and food safety research regarding insects as a new food

42 matrix were initiated, e.g. by EFSA1. In the past five years, many basic microbiological food

43 safety questions were formulated and investigated, as reviewed by Garofalo et al. (2019)2. In

44 summary, freshly reared insects may be highly contaminated with diverse microorganisms.

45 Even though insects are not likely to be consumed in a raw status, it is important to know which

46 organisms should be targeted by heat and other processing treatments. It was demonstrated both

47 for fungi and bacteria for yellow mealworms (Tenebrio molitor)3, lesser mealworms

48 (Alphitobius diaperinus)4 and tropical house crickets (Gryllodes sigillatus)5 that high numbers

49 of these microorganisms can easily be reduced, e.g. by applying a mild heat treatment..

50 However, it was repeatedly observed that a fraction of the bacteria, i.e. the endospores, can

51 survive those mild heat treatments3,6. Bacillus and Clostridium are two relevant bacterial genera

52 that are able to produce endospores. Since several members of these genera are food pathogens

53 (e.g. B. cereus, C. perfringens, C. botulinum), the survival of endospores in the insect matrix

54 may pose food safety risks. DNA-based studies have already identified, among others, the

55 spore-forming genera Bacillus, Brevibacillus, Clostridium, Lysinibacillus and Paenibacillus in

56 edible insects and insect products produced in Europe7–12. Yet, it is generally not clear which

57 specific species were found and whether or not they were present in spore form. Consequently,

58 the first goal of this research (experiment 1) was to identify which bacterial endospores can be

59 present in raw samples of the two edible insect species reared mostly for human consumption 3

60 nowadays: the yellow mealworm (Tenebrio molitor) and the house cricket (Acheta domesticus).

61 Specific attention was given to B. cereus-related isolates, which were further characterised by

62 exploring specific genes with qPCR.

63 Another aspect that has not been explored so far is the presence of pathogenic foodborne

64 viruses in insects for food. The possibility that foodborne viruses such as hepatitis A virus

65 (HAV), hepatitis E virus (HEV) or norovirus (NoV) can be present in insects or derived

66 foodstuffs has been proposed by EFSA already in 20151. The three said viruses are the most

67 commonly transmitted viruses through foodstuffs13 and are relevant to be investigated for

68 insects and insect-based food matrices. This research was the first to elucidate whether the

69 selected human foodborne viruses HAV, HEV and NoV genogroup II (GII, responsible for the

70 majority of norovirus cases14) are present in raw mealworm and cricket species, and if so, in

71 what quantity (experiment 2).

72 When studying the microbiota of insects, it is very plausible that the organisms present

73 in their gut and on their exoskeleton differ between insects cultivated in the laboratory and the

74 same species reared at industrial scale, where the hygiene levels may be lower and hence the

75 house flora different. Since the aim of this work was to study bacterial endospores and viruses

76 with respect to safety for the consumer, insect samples obtained from several commercial insect

77 producers were studied and multiple batches per producer were included in the experimental

78 set-up.

80 Results and discussion

81 Bacterial endospore counts

82 In experiment 1, the raw insect samples investigated were first subjected to bacterial

83 endospore counts and, as a comparison, total viable counts (Figure 1). The obtained values for 4

84 total viable aerobic (between 7.5 and 8.7 log cfu/g) and anaerobic (7.4 to 8.7 log cfu/g) counts

85 are similar to the typically high counts for raw edible insects reported earlier2,15. Additionally,

86 the values for the aerobic and anaerobic counts are highly comparable. This may suggest that a

87 large fraction of both counts exists of facultative anaerobic species. When comparing both

88 insect species investigated, it is clear that the total viable counts for house crickets are higher

89 than those for yellow mealworms. This confirms what was proposed earlier16, being that

90 microbiological findings of particular insect species should not be generalised for all edible

91 insects, as nevertheless is done in literature occasionally. For bacterial endospore counts, which

92 represent the fraction of the total viable counts present as spore form significant variation was

93 observed between samples from the same insect species, compared to the total viable counts.

94 This variation may result from differences in (hygiene in) the rearing environment and/or feed,

95 and is commonplace for insects2,9,11. Again, anaerobic and aerobic endospore counts displayed

96 comparable trends (Figure 1).

97 Currently in the insect sector, edible insects are typically subjected to a heat treatment

98 prior to processing and consumption, such as boiling and/or oven drying. It has been

99 demonstrated3,5,6 that bacterial endospores can survive these treatments and remain present and

100 viable in insects for human consumption. The highest endospore count observed in this study

101 was 4.2 log cfu/g (for a house cricket sample). While no European microbiological criteria exist

102 specifically for edible insects, this count does surpass the lower action limit for Bacillus cereus

103 in edible insects as published by the Belgian Federal Agency for the Safety of the Food Chain17.

104 In the case that a large fraction of these spores found would consist of B. cereus, this would

105 pose a health risk for human consumption. For clostridia, no criteria or action limits exist, but

106 it is advised to pay attention to these potential pathogens as well.

107

108 Identification of bacterial endospores

109 Microbiological health risks of food products are determined not only by population

110 densities of microbial (sub)groups, but also by the particular species present in the foodstuff. In

111 this study, we focussed on bacterial endospores, which are resistant against several treatments

112 that are harmful for vegetative cells. Pathogenic bacteria present in their spore form are of major

113 concern for edible insects.

114 In total, 142 endospore-forming isolates were collected from the bacterial endospore

115 count plates obtained from yellow mealworm and house cricket samples. These isolates were

116 present as endospore in the raw insect matrix and survived the pasteurisation treatment during

117 analysis (as part of the ISO-based endospore count), germinated and grew on Plate Count Agar.

118 By homogenising the sample prior to the isolation of endospores, it was possible to obtain a

119 complete view on the presence of endospores both in and on the surface of the insect16. A

120 complete view is necessary, since the insects considered are processed and consumed

121 completely. On the other hand, the approach used in this study impedes the possibility to locate

122 the microorganisms on or in the insect body. Likewise, it is not possible to track the origin of

123 the microorganisms, i.e. whether they were introduced via the substrate, via the rearing

124 environment or during packaging or transport.

125 The isolates collected were assigned to four different genera of endospore-forming

126 bacteria, including Bacillus, Lysinibacillus, Brevibacillus and Clostridium. Based on further

127 identification, the Bacillus genus was further divided into Bacillus cereus group (also B. cereus

128 sensu lato (s.l.)) and Bacillus (non-cereus) (Figure 2A; Supplementary Table 1). Of the 50

129 isolates obtained from yellow mealworms, 20 (i.e. 40%) were identified as a member of the

130 Bacillus cereus group. For house crickets, even 79 out of 92 isolates (86%) were identified as

131 B. cereus s.l. This may indicate that house crickets are highly susceptible to colonisation with

132 B. cereus group members.

133 The B. cereus group consists, among a few recently described species, of the closely

134 related bacteria B. cereus sensu stricto, B. thuringiensis, B. weihenstephanensis, B. wiedmannii,

135 B. anthracis, B. mycoides, B. pseudomycoides, B. cytotoxicus, and B. toyonensis18–21. Most

136 isolates assigned to the B. cereus group were identified as B. cereus s.s. or B. paramycoides

137 (proposed novel member22). Due to the close genetic relationships within the B. cereus group,

138 however, it is hard to distinguish between the different species compiled in this group based on

139 the 16 S rRNA gene21. Several species of the B. cereus group are relevant with respect to human

140 health, agriculture or food safety. In the case of edible insects, presence of the well-known

141 pathogen B. cereus s.s. may pose a severe food safety risk for consumers, but also some B.

142 weihenstephanensis and B. cytotoxicus strains were reported as potential human (foodborne)

143 pathogens23,24. For insect rearing, B. thuringiensis may be harmful because of its insecticidal

144 properties25. Whatever B. cereus group species were present in the insects, the B. cereus group

145 poses an important threat for the edible insect sector and should be monitored. Also, given the

146 fact that the B. cereus group contains psychrotrophic strains20,26, chilled preservation of heat-

147 treated insects can still allow the growth (and toxin production) of food pathogens such as B.

148 cereus s.s. Mitigation strategies to lower the risks regarding pathogenic bacterial endospores

149 include (i) the use of substrates that do not contain these species, which requires a thorough

150 quality control, (ii) decontamination of the insects via the hurdle or combination strategy (e.g.

151 steam, combining heat with pressure, pulsed electric fields, irradiation, …) or (iii) prevent

152 spores from germinating (e.g. by reducing the water activity, pH and/or temperature of the

153 matrix, …).

154 A few previous studies on the microbiota harboured by edible insects also report

155 members of the B. cereus group. For example, B. cereus s.s. and B. weihenstephanensis were

156 detected by PCR-DGGE (Denaturing Gradient Gel Electrophoresis) in edible processed insects

157 bought in Belgium and the Netherlands10. In another study11, processed edible insects were

158 described to contain up to 6.6 log cfu/g presumptive B. cereus, of which B. cytotoxicus, B.

159 cereus s.s. and B. thuringiensis isolates could be identified using sequencing and biomolecular

160 identification.

161 The second most observed endospore-forming bacteria were members of the genus

162 Lysinibacillus (21 out of 142 isolates, or 15%; Figure 2A) and the majority was identified as L.

163 fusiformis (Supplementary Table 1). This spore former, renamed from Bacillus fusiformis27,

164 rarely acts as a human pathogen, but especially its relative L. sphaericus has been reported as a

165 potent insect pathogen. L. sphaericus preferably targets mosquitos, but activity against other

166 insects species has been observed as well28. Furthermore, Lysinibacillus spp. were described as

167 potential proteolytic spoilage organisms that can thus alter product quality29,30. Also

168 previously11, Lysinibacillus sp. was detected in edible insects.

169 Besides B. cereus members, also other, non-cereus Bacillus spp. were detected (16/142

170 isolates, 11%; Figure 2A). The genus Bacillus has regularly been found in edible

171 insects6,10,11,31,32, but it has rarely been described at species level. According to the identification

172 results (Supplementary Table 1), non-cereus Bacillus isolates may correspond to several species

173 including B. pumilus, B. altitudinis, B. siamensis, Bacillus licheniformis, B. vallismortis and B.

174 subtilis. The latter three species are closely related and, comparably to the B. cereus group, they

175 are members of the B. subtilis group33. As was also the case for the B. cereus group, the 16 S

176 rRNA gene does not allow for species-level identification for B. subtilis group members33. B.

177 subtilis and its relatives are generally considered as spoilage organisms rather than human food

178 pathogens34, but were linked to toxin production and a few cases of foodborne illness in the

179 past35,36. On the other hand, they are also frequently described as beneficial for plants and/or

180 animals e.g. as biocontrol organism or probiotic37,38, which may pose opportunities for

181 industrial valorisation, also in the insect sector.

182 Six isolates (3%; Figure 2A) were ascribed to the genus Brevibacillus and identified as

183 either B. laterosporus or B. halotolerans (Supplementary Table 1). They were only obtained

184 from yellow mealworms and not from house crickets. In previous research7, a Brevibacillus sp.

185 was detected by Illumina sequencing in yellow mealworms, in abundances up to 28%. B.

186 laterosporus has been reported as insect pathogen39. Together with B. thuringiensis, a member

187 of the B. cereus group, and Lysinibacillus sphaericus, a realistic risk for edible insect rearing

188 exists when these entomopathogenic bacteria are present. Further, Brevibacillus spp. can act as

189 spoilage organism and impact the quality of insect food products29. On the other hand, B.

190 laterosporus has also been described as a broad-spectrum antimicrobial species against

191 phytopathogenic bacteria and fungi39, which may involve agricultural benefits.

192 While both aerobic and anaerobic bacterial endospore counts were performed and

193 isolates were picked randomly from all plates from all samples, it is striking that almost solely

194 aerobic and/or facultative anaerobic organisms were identified. (Strict) anaerobic spore formers

195 such as specific Clostridium spp. were encountered in edible insects before, including yellow

196 mealworms and house crickets7–9,40. In this study, however, only one isolated was assigned to

197 the genus Clostridium, yet, with a low sequence identity (69.4 %, Supplementary Table 1). The

198 genus Clostridium is of concern regarding food safety, since it contains the food pathogens C.

199 perfringens and C. botulinum. They may both produce potent toxins that cause gastroenteris

200 and botulism, respectively. Food poisoning caused by C. perfringens, however, generally

201 requires the ingestion of 108 cells that can then sporulate (and release their toxin) in the

202 intestine41. The total viable anaerobic counts observed in this study may exceed 8 log cfu/g, but

203 probably do not consist solely of Clostridium cells. A selective Clostridium count may be useful

204 to confirm this. All previous observations of (possible) anaerobic spore-forming bacteria were,

205 however, based on culture-independent methods. The random selection of isolates and/or the

206 strict anaerobic conditions necessary for cultivating certain Clostridium species may have

207 impeded to obtain anaerobic species that might have been present in the samples investigated.

208 Accordingly, the specific isolation and identification of anaerobic bacterial spore formers would

209 form an interesting addition to this research.

210

211 qPCR detection of B. cereus-related genes

212 It has been demonstrated that the 16 S rRNA gene is extremely similar (97.34 to 100%

213 inter-species similarity) between all B. cereus group members 20,42. The 16 S rRNA gene-based

214 identification method applied in this study was therefore complemented by studying additional

215 genetic markers to further characterise and discriminate the B. cereus s.l. isolates. Firstly, all B.

216 cereus s.l. were subjected to a qPCR assay targeting the panC gene encoding a pantothenate-β-

217 alanine ligase43. The presence of this gene was used to confirm the assignment of the isolates

218 to the B. cereus group, since panC genotyping is a typical method to group B. cereus s.l. species

219 based on phenotypic characteristics44. Next, by targeting the cesA gene45, each isolate was

220 assessed for plasmid presence encoding for the production of the heat-resistant cereulide toxin,

221 the most severe food safety risk46,47. This allows to evaluate the possibility for each isolate to

222 act as a human pathogen of the emetic type, regardless of its taxonomic classification.

223 Additionally, to date, this virulence plasmid was only encountered in B. cereus s.s. and B.

224 weihenstephanensis, while it was reported not to be present in a large collection of B.

225 thuringiensis, B. mycoides and B. pseudomycoides strains18,47. In the (edible) insect sector, the

226 distinction between foodborne pathogen (B. cereus s.s. or B. weihenstephanensis) and insect-

227 borne pathogen (B. thuringiensis) is of major interest, and presence or absence of the cesA gene

228 provides additional information in this regard.

229 Of the 98 B. cereus group isolates assessed for the panC gene, only three were found to

230 be negative. Consequently, the identification of 95 B. cereus group isolates was considered

231 reliable. Next, in 64 of the 98 isolates (65 %), the cesA gene was detected (Figure 2B). This

232 indicates that those isolates may be able to produce the cereulide toxin under suitable

233 conditions. Influencing factors are for instance temperature, the food matrix (e.g. water activity

48,49 234 (aw)) and the number of B. cereus cells present . While the precise impact of water activity

235 is not fully understood yet, a minimal aw-value of 0.953 has been proposed for the production

236 of cereulide50. The water activity of raw edible mealworms and crickets, as were used in this

16 237 study, ranges between 0.95 and 0.98 and is consequently suitable for cereulide production.

238 The amount of B. cereus required to be present for toxin production to take place is typically

239 reported as 3 to 5 log cfu/g18. B. cereus counts were not assessed in this study, but in literature,

240 presumptive B. cereus were already detected up to 6.6 log cfu/g in processed edible mole

241 crickets11.

242 The presence of the cereulide plasmid in B. cereus in its endospore form poses an

243 increased food safety risk. The number of bacterial endospores in the samples investigated was

244 maximally 4.2 log cfu/g and did not relate solely to B. cereus s.l. Despite this observation, a

245 risk is still present because after a heat treatment (which is survived by the spores and can also

246 even activate them), the endospores can germinate in the insect matrix, and in absence of

247 competition, multiply and produce cereulide. A second heating step will not be sufficient to

248 destroy the toxin and thus the risk. Consequently, the exact conditions for and the extent of

249 cereulide production that can occur in edible insects and derived food products is a future

250 research need.

251

252 Presence of foodborne viruses

253 Three different RT-qPCR assays were optimised to detect and quantify the RNA viruses

254 HAV, HEV and NoV GII in insect RNA extracts and assessed for their reliability. The addition

255 of an Internal Amplification Control (IAC) to the RT-qPCR assay has proven to be a robust

256 control in virus analysis in food matrices51 and has become common practice52,53. Similarly, in

257 this study, IACs were included in all samples investigated and were systematically recovered

258 in the three assays (Ct values ± 35), while negative IAC controls showed no amplification after

259 40 cycles, proving the reactions to be successful. In order to quantify the viral titer, standard

260 curves were constructed and showed to be of good quality, according to their calculated

261 parameters R², slope and efficiency. R² was determined at 0.998 for HAV and 0.999 for HEV

262 and NoV GII. Next, the slope was calculated as -3.737, -3.287 and -3.625 for HAV, HEV and

263 NoV GII, respectively. Finally the efficiency showed values of 85.192%, 101.487% and

264 88.739% for HAV, HEV and NoV GII respectively. Based on the recovery of the reference

265 virus cDNA fragments, the detection limit for each assay was set at 100 copies/µl undiluted

266 RNA extract. Also the consistent amplification of positive controls (Ct values ± 31) and absence

267 of qPCR signal for negative controls contributed to the conclusion that the RT-qPCR assays

268 were reliable.

269 In none of the samples any of the viruses HAV, HEV or NoV GII could be detected,

270 meaning that the viruses were either not present or the samples contained less than 100 viral

271 RNA copies/µl extract. Consequently, the transmission risk for HAV, HEV and NoV GII via

272 yellow and lesser mealworm and (tropical) house cricket to humans is low. However, other

273 insect species and/or other foodborne viruses should still be investigated in future research, and

274 it should involve industrial samples.

275

276 Conclusions

277 In an extensive collection of industrially produced edible insects, this study uncovers a

278 large contrast between the high risk for foodborne illnesses originating from bacterial spores

279 and the low risk from human viruses in edible mealworms and crickets. For the bacterial

280 endospores, the largest fraction of isolates was identified as members of the Bacillus cereus

281 group, which contains foodborne and insect-borne pathogens. Sixty-five % of the B. cereus

282 group isolates contained the cereulide plasmid that can allow the production of a heat-resistant

283 toxin. The presence of the insect pathogens Brevibacillus laterosporus and Lysinibacillus sp.,

284 on the other hand, may involve economic risks for the insect rearer due to production losses.

285 Altogether, the possibility that batches are rejected due to microbiological problems may

286 increase, with concomitantly an increasing economic impact, since the sector is growing as

287 outlined above. Appropriate mitigation strategies to control the presence and outgrowth of

288 bacterial endospores should hence be adopted. In literature, limited information is available,

289 but insect producing and processing companies pay efforts in this field.

290 The presence of foodborne viruses in edible insects was investigated here for the first

291 time. Using methods shown to be reliable, hepatitis A virus, hepatitis E virus and norovirus

292 genogroup II were not detected (detection limit of 100 copies/µl extract). Hence, food safety

293 risks related to these viruses can be stated to be very low. However, as industrial insect

294 production methods are evolving due to automation and upscaling, virus detection is advised to

295 be continued and also performed for other species than those considered in this study.

296 13

297 Methods

298 Sample collection and preparation

299 For experiment 1 (identification of bacterial endospores in edible insects), four samples

300 of living yellow mealworms (Tenebrio molitor) and four samples of living house crickets

301 (Acheta domesticus) were collected at the end of their rearing cycle, without further processing.

302 For each insect species, two different industrial rearing companies in Belgium (yellow

303 mealworms) or Belgium and the Netherlands (house crickets) were sampled two times each (2

304 samples x 2 rearers x 2 sampling moments or batches). All living insect samples were packed

305 by the producer in a closed tray or box and transported at ambient temperature.

306 For experiment 2 (detection of foodborne viruses in edible insects), the same 17

307 untreated (raw) mealworm and cricket samples that were investigated previously7,16 for their

308 bacterial composition (yellow mealworm, house cricket and tropical house cricket (Gryllodes

309 sigillatus) samples from different rearing companies and batches) were employed, as well as an

310 extra sample of lesser mealworms (Alphitobius diaperinus) obtained in another study4 where it

311 was used to study microbial dynamics during rearing (larvae day 35, post-harvest).

312 Samples were treated aseptically as from arrival in the laboratory. Prior to the analyses,

313 dead specimens were removed with sterile tweezers and remaining insects were sedated by

314 cooling them for approximately 1 h at 4 °C. Subsequently, the samples were homogenised by

315 pulverisation with a sterilised hand-held mixer (Bosch CNHR 25, Belgium) as described

316 earlier9. This way, the whole insect microbiota can be considered.

317

318 Bacterial endospore counts and isolation

319 All samples employed in experiment 1 were subjected to aerobic and anaerobic bacterial

320 endospore counts in threefold. Following the procedure described earlier16, for each analysis, 5 14

321 g of pulverised insects were diluted in 45 g peptone physiological salt solution (0.85% NaCl,

322 0.1% peptone, Biokar Diagnostics, Beauvais, France) and homogenised in a Bagmixer®

323 (Interscience, Saint Nom, France) for 1 min. This primary dilution was given a pasteurisation

324 treatment at 80 °C for 10 min and was then further diluted and plated on Plate Count Agar

325 (PCA, Biokar diagnostics) using the spread plate technique to easily pick colonies later. Both

326 aerobic and anaerobic endospore counts were determined after incubation for 48 h at 37 °C54.

327 As a comparison, a similar but unpasteurised dilution series was used to determine the total

328 viable aerobic and anaerobic counts of the samples (pour plate technique, PCA, 72 h at 30 °C).

329 Anaerobic conditions were generated in Anaerocult containers (2.5 L, VWR International,

330 Leuven, Belgium) using ‘AnaeroGen 2.5 L atmosphere generation systems’ (Thermo Fisher

331 Scientific, Asse, Belgium) and evaluated using resazurin indicators (BR0055N, Thermo Fisher

332 Scientific).

333 For each sample, after incubation, several colonies with various morphologies were

334 picked from the countable plates and streaked on nutrient agar (NA, Biokar Diagnostics, 24 h

335 at 37 °C, (an)aerobic incubation depending on their origin) to form axenic cultures. An

336 individual colony was subsequently incubated overnight in nutrient broth (NB, Biokar

337 Diagnostics, 18 h at 37 °C) and stored at - 80 °C after addition of glycerol to a final

338 concentration of 50 % (v/v). This way, in total 67 and 95 spore-forming isolates were collected

339 from the yellow mealworm and house cricket samples, respectively.

340

341 Identification of bacterial endospores

342 All 162 endospore isolates were, after recultivation in NB, subjected to

343 phenol/chlorophorm genomic DNA extraction as described earlier55 and subsequently to PCR

344 (T100™ Thermal Cycler, Bio-Rad, Temse, Belgium), amplifying the 16 S ribosomal RNA

345 (rRNA) gene. The PCR reaction (20 µl) contained 1.25 U of TaKaTa ExTaq Polymerase and 1

346 x ExTaq Buffer (Clontech Laboratories, Palo Alto, CA, USA), 312.5 µM of each dNTP, 1.0

347 µM of each primer (27F and 1492R, Table 1), and 5 ng genomic DNA (measured by a Nanodrop

348 spectrophotometer, Thermo Fisher Scientific). PCR conditions included initial denaturation for

349 2 min at 95 °C, followed by 34 cycles of 45 s denaturation at 95 °C, 45 s annealing at 59 °C

350 and 45 s elongation at 72 °C, and final elongation for 10 min at 72 °C. Obtained amplicons

351 were sequenced using the same reverse primer as used in the PCR (1492R, Table 1) by

352 Macrogen Europe (Amsterdam, the Netherlands). Resulting sequences were aligned and

353 trimmed to an average read length of 861 bp. Isolates were subsequently classified using the

354 EzBioCloud 16S rRNA gene database56. This way, a total of 142 isolates (50 from yellow

355 mealworms, 92 from house crickets) could be identified.

356

357 qPCR detection of B. cereus-related genes panC and cesA

358 For 99 isolates classified as B. cereus s.l., two additional genes were traced using qPCR.

359 For each extract, firstly, a qPCR assay targeting the panC gene was applied in order to confirm

360 the association with B. cereus s.l. Next, presence of the cereulide toxin pathway was

361 investigated by assessing the cesA gene. All qPCR reactions (10 µl) were executed in a

362 QuantStudio 3 qPCR system (Applied Biosystems, Thermo Fisher Scientific) and contained 5

363 µl PowerUp SYBR Green Master Mix with UNG (uracil-N glycosylase to prevent carryover

364 contamination; Thermo Fisher Scientific), 1 µM (panC) or 0.4 µM (cesA) of each primer (Table

365 1) and 60 ng of DNA (measured by a MySpec spectrophotometer, VWR International). qPCR

366 reaction conditions for the panC detection followed the master mix supplier’s protocol: 2 min

367 at 50 °C (UNG activation) and 2 min of initial denaturation at 95 °C, followed by 40 cycles of

368 15 s denaturation at 95 °C and combined annealing and elongation for 1 min at 60 °C. For the

369 cesA detection, the same conditions were applied, except for the initial denaturation step which

370 was shortened to 20 seconds57. Each qPCR was followed by a melt curve analysis by raising

371 the temperature gradually from 60 to 95 °C (0.1 °C/s) while continuously monitoring

372 fluorescence. Positive (B. cereus emetic type, LMG 17603, Belgian Coordinated Collections of

373 Microorganisms (BCCM)) and negative (no template, sterile ultra-pure water instead) controls

374 were included. Data analysis was executed using the online application “Design and Analysis”

375 in the qPCR DataConnect cloud software (Thermo Fisher Scientific). Successful detection was

376 considered when the melt curve analysis revealed a single peak at the expected melting

377 temperature, and when obtained Ct values did not exceed these of the negative control (33 for

378 cesA and 35 for panC).

379

380 RT-qPCR quantification of selected viruses

381 In order to detect and quantify HAV, HEV and NoV GII, previously described

382 methods52 were optimised and used to screen the insect samples in experiment 2. Standard

383 quantification curves and Internal Amplification Controls (IACs)51 were included to assess

384 quantification and reliability. Corresponding virus reference cDNA was used as standard, either

385 obtained from an RNA transcript (NoV GII after RT-PCR, see below) or plasmids containing

386 virus genome fragments (HAV and HEV after PCR targeting the reference fragment). IAC

387 RNAs were generated by transforming IAC-containing plasmids into chemically competent

388 Escherichia coli cells (One Shot™ TOP10, Thermo Fisher Scientific) and transcription to RNA

389 using the T7 High Yield Transcription kit (Thermo Fisher Scientific).

390 Homogenised insect samples (experiment 2) were subjected in twofold to a total RNA

391 extraction, using the Mo Bio RNA PowerSoil kit (manufacturer’s protocol, Carlsbad, CA,

392 USA). Next, each RNA extract was subjected to a two-step reverse transcription real-time PCR

393 (RT-qPCR) assay for each target virus separately. First, a tenfold RNA extract dilution was

394 used to translate target virus RNA into cDNA (Table 1). In this step, the IAC corresponding to

395 the target virus was added to exclude false negative results. Each reaction (20 µl) was performed

396 using the High Capacity cDNA Reverse Transcription kit (Applied Biosystems, Thermo Fisher

397 Scientific) and contained 2.0 µl RT Buffer, 1.0 µl MultiScribe™ Reverse Transcriptase, 2 µM

398 reverse primer (Table 1), 0.8 µl dNTP Mix (100 mM), 0.5 µl RNase Inhibitor, 1000 IAC copies,

399 and 40 ng RNA (as measured by MySpec spectrophotometer). Reaction conditions (iCycler

400 Thermal Cycler, Bio-Rad) were as described by the supplier: 10 min at 25 °C, 2 h at 37 °C, and

401 5 min at 85 °C. Positive (108 IAC copies) and negative (no template, sterile ultra-pure water

402 instead) controls were included for all virus assays. Next, for each sample, the obtained cDNA

403 was used as a template in the subsequent qPCR which aimed to detect and quantify HAV, HEV

404 or NoV GII as well as to detect the corresponding IAC at the same time. The three virus assays

405 were conducted based on protocols described earlier52. All analyses were performed in

406 duplicate and each reaction (20 µl) contained 1x TaqMan Universal Master Mix with UNG

407 (Applied Biosystems, Thermo Fisher Scientific), 0.25 µM of each primer (Table 1), 0.25 µM

408 virus probe (FAM-labelled, Table 1) and 0.10 µM IAC probe (VIC-labelled, Table 1). All

409 reactions were executed using a QuantStudio 3 qPCR system. qPCR conditions were as follows:

410 2 min at 50 °C, 10 min at 95 °C and 40 cycles of 15 s denaturation at 95 °C and 60 s combined

411 annealing and elongation at 60 °C. Virus quantities and standard curve parameters were

412 calculated in the qPCR DataConnect cloud software. Again, positive (1000 virus cDNA copies)

413 and negative (no template, sterile ultra-pure water instead) controls were included. Only when

414 the IAC was detected below the negative control Ct, the RT-qPCR assay was considered

415 successful.

416

417 Statistical analyses

418 To determine differences in microbial counts between samples in experiment 1, results

419 were subjected to one-way ANOVA with all-pairs Tukey-Kramer HSD post hoc test. Normality

420 and homoscedasticity assumptions were tested for the analysed data sets using the Shapiro-

421 Wilk W test and Levene’s test, respectively. In the cases that homoscedasticity was not

422 confirmed, Welch’s ANOVA was used instead. Results were displayed as a Mean Diamonds

423 plots with proportional x-axis. JMP pro 14.0.0 (SAS Institute Inc., Cary, NC, USA) was used

424 with significance level 0.05 for the statistical tests and creation of the plots.

425

426 Data availability

427 Sequencing data obtained in this study have been deposited in GenBank (National

428 Center for Biotechnology Information, NCBI) with the accession codes MN508485 to

429 MN508527 (Supplementary Table 1). All other relevant data are available from the

430 corresponding author on request.

431

432 Acknowledgements

433 Virus references and IACs were kindly provided by I. Di Bartolo from the Italian Istituto

434 Superiore di Sanità (ISS, Rome, Italy) and N. Cook from Fera Science Ltd. (York, United

435 Kingdom). Additionally, the authors would like to thank A. Paeleman (Scientia Terrae Research

436 Institute, Sint-Katelijne-Waver, Belgium) for her expertise and assistance in designing and

437 optimising the qPCR protocols, S. Crauwels (KU Leuven, Leuven, Belgium) for processing the

438 sequencing results and R. Smets (KU Leuven, Geel) for the statistical help. Also J. Franciotti,

439 L. De Vrindt, N. Huybrechts, M. Gerits, S. Machtajiw, J. Plas and E. Van Vossole are

440 acknowledged for their assistance in the lab. This research was financially supported by 19

441 Flanders Innovation & Entrepreneurship (VLAIO) [Project 141129] as well as Internal Funds

442 KU Leuven [PDM/18/159].

443

444 Author Contributions

445 D.V. designed, prepared and executed all experiments, including sample collection and

446 preparation, microbiological analyses, DNA and RNA extractions and PCR and RT-qPCR

447 reactions. D.V. analysed the data, constructed tables and figures and wrote the main text. B.L.

448 and L.V. supervised the study, provided additional insight in data analysis and revised the

449 manuscript.

450

451 Competing Interests statement

452 The authors declare no competing interests.

453

454 References

455 1. EFSA Scientific Committee. Risk profile related to production and consumption of

456 insects as food and feed. EFSA J. 2015 13, 60 pp. (2015).

457 2. Garofalo, C. et al. Current knowledge on the microbiota of edible insects intended for

458 human consumption: A state-of-the-art review. Food Res. Int. 125, 108527 (2019).

459 3. Vandeweyer, D., Lenaerts, S., Callens, A. & Van Campenhout, L. Effect of blanching

460 followed by refrigerated storage or industrial microwave drying on the microbial load

461 of yellow mealworm larvae (Tenebrio molitor). Food Control 71, 311–314 (2017).

462 4. Wynants, E. et al. Microbial dynamics during production of lesser mealworms

463 (Alphitobius diaperinus) for human consumption at industrial scale. Food Microbiol.

464 70, 181–191 (2018).

465 5. Vandeweyer, D. et al. Microbial Dynamics during Industrial Rearing, Processing, and

466 Storage of Tropical House Crickets (Gryllodes sigillatus) for Human Consumption.

467 Appl. Environ. Microbiol. 84, AEM.00255-18 (2018).

468 6. Klunder, H. C., Wolkers-Rooijackers, J., Korpela, J. M. & Nout, M. J. R.

469 Microbiological aspects of processing and storage of edible insects. Food Control 26,

470 628–631 (2012).

471 7. Vandeweyer, D., Crauwels, S., Lievens, B. & Van Campenhout, L. Metagenetic

472 analysis of the bacterial communities of edible insects from diverse production cycles

473 at industrial rearing companies. Int. J. Food Microbiol. 261, 11–18 (2017).

474 8. Garofalo, C. et al. The microbiota of marketed processed edible insects as revealed by

475 high-throughput sequencing. Food Microbiol. 62, 15–22 (2017).

476 9. Stoops, J. et al. Microbial community assessment of mealworm larvae (Tenebrio

477 molitor) and grasshoppers (Locusta migratoria migratorioides) sold for human

478 consumption. Food Microbiol. 53, Part B, 122–127 (2016).

479 10. Osimani, A. et al. Insight into the proximate composition and microbial diversity of

480 edible insects marketed in the European Union. Eur. Food Res. Technol. 243, 1157–

481 1171 (2017).

482 11. Fasolato, L. et al. Edible processed insects from e-commerce: Food safety with a focus

483 on the Bacillus cereus group. Food Microbiol. 76, 296–303 (2018).

484 12. Stoops, J. et al. Minced meat-like products from mealworm larvae (Tenebrio molitor

485 and Alphitobius diaperinus): Microbial dynamics during production and storage. Innov.

486 Food Sci. Emerg. Technol. 41, 1–9 (2017).

487 13. Petrović, T. & D’Agostino, M. Viral Contamination of Food. in Antimicrobial Food

488 Packaging (ed. Barros-Velázquez, J.) 65–79 (Academic Press, 2016).

489 doi:10.1016/B978-0-12-800723-5.00005-X

490 14. Da Silva, A. K. et al. Evaluation of removal of noroviruses during wastewater

491 treatment, using real-time reverse transcription-PCR: Different behaviors of

492 genogroups I and II. Appl. Environ. Microbiol. 73, 7891–7897 (2007).

493 15. De Smet, J. et al. Stability assessment and laboratory scale fermentation of pastes

494 produced on a pilot scale from mealworms (Tenebrio molitor). LWT 102, 113–121

495 (2019).

496 16. Vandeweyer, D., Crauwels, S., Lievens, B. & Van Campenhout, L. Microbial counts of

497 mealworm larvae (Tenebrio molitor) and crickets (Acheta domesticus and Gryllodes

498 sigillatus) from different rearing companies and different production batches. Int. J.

499 Food Microbiol. 242, 13–18 (2017).

500 17. FASFC. Appendix: Table with regulatory criteria and action limits. Action limits for

501 microbiological contaminants in foods (2019). Available at:

502 http://www.favv.be/thematischepublicaties/_documents/Narval_microbio_VCT_JWS_

503 ADK__CKS_27_feb_2019_nl.xlsx.

504 18. EFSA Panel on Biological Hazards. Risks for public health related to the presence of

505 Bacillus cereus and other Bacillus spp. including Bacillus thuringiensis in foodstuffs.

506 EFSA J. 14, 93 (2016).

507 19. Rasigade, J. P., Hollandt, F. & Wirth, T. Genes under positive selection in the core

508 genome of pathogenic Bacillus cereus group members. Infect. Genet. Evol. 65, 55–64

509 (2018).

510 20. Liu, Y. et al. Genomic insights into the taxonomic status of the Bacillus cereus group.

511 Sci. Rep. 5, 14082 (2015).

512 21. Ehling-Schulz, M., Lereclus, D. & Koehler, T. M. The Bacillus cereus Group: Bacillus

513 Species with Pathogenic Potential. Microbiol. Spectr. 7, (2019).

514 22. Liu, Y. et al. Proposal of nine novel species of the Bacillus cereus group. Int. J. Syst.

515 Evol. Microbiol. 67, 2499–2508 (2017).

516 23. Guinebretière, M.-H. et al. Bacillus cytotoxicus sp. nov. is a novel thermotolerant

517 species of the Bacillus cereus group occasionally associated with food poisoning. Int. J.

518 Syst. Evol. Microbiol. 63, 31–40 (2013).

519 24. Stenfors, L. P., Mayr, R., Scherer, S. & Granum, E. Pathogenic potential of fifty

520 Bacillus weihenstephanensis strains. FEMS Microbiol. Lett. 215, 47–51 (2002).

521 25. Bravo, A., Gill, S. S. & Soberón, M. Mode of action of Bacillus thuringiensis Cry and

522 Cyt toxins and their potential for insect control. Toxicon 49, 423–435 (2007).

523 26. Martínez, S., Borrajo, R., Franco, I. & Carballo, J. Effect of environmental parameters

524 on growth kinetics of Bacillus cereus (ATCC 7004) after mild heat treatment. Int. J.

525 Food Microbiol. 117, 223–7 (2007).

526 27. Ahmed, I., Yokota, A., Yamazoe, A. & Fujiwara, T. Proposal of Lysinibacillus

527 boronitolerans gen. nov. sp. nov., and transfer of Bacillus fusiformis to Lysinibacillus

528 fusiformis comb. nov. and Bacillus sphaericus to Lysinibacillus sphaericus comb. nov.

529 Int. J. Syst. Evol. Microbiol. 57, 1117–1125 (2007).

530 28. Berry, C. The bacterium, Lysinibacillus sphaericus, as an insect pathogen. J. Invertebr.

531 Pathol. 109, 1–10 (2012).

532 29. Lücking, G., Stoeckel, M., Atamer, Z., Hinrichs, J. & Ehling-Schulz, M.

533 Characterization of aerobic spore-forming bacteria associated with industrial dairy

534 processing environments and product spoilage. Int. J. Food Microbiol. 166, 270–279

535 (2013).

536 30. De Jonghe, V. et al. Toxinogenic and spoilage potential of aerobic spore-formers

537 isolated from raw milk. Int. J. Food Microbiol. 136, 318–325 (2010).

538 31. Wang, Y. & Zhang, Y. Investigation of gut-associated bacteria in Tenebrio molitor

539 (Coleoptera: Tenebrionidae) larvae using culture-dependent and DGGE methods. Ann.

540 Entomol. Soc. Am. 108, 941–949 (2015).

541 32. Osimani, A. et al. The bacterial biota of laboratory-reared edible mealworms (Tenebrio

542 molitor L.): From feed to frass. Int. J. Food Microbiol. 272, 49–60 (2018).

543 33. Wang, L. T., Lee, F. L., Tai, C. J. & Kasai, H. Comparison of gyrB gene sequences,

544 16S rRNA gene sequences and DNA-DNA hybridization in the Bacillus subtilis group.

545 Int. J. Syst. Evol. Microbiol. 57, 1846–1850 (2007).

546 34. De Vos, P. et al. Bergey’s Manual of Systematic Bacteriology Volume 3. (Springer

547 Science+Business Media, 2009). doi:10.1007/978-0-387-68489-5

548 35. Kramer, J. M. & Gilbert, R. J. Bacillus cereus and other Bacillus species. in Foodborne

549 Bacterial Pathogens (ed. Doyle, M.) 21–70 (Marcel Dekker, 1989).

550 36. From, C., Pukall, R., Schumann, P., Hormazábal, V. & Granum, P. E. Toxin-producing

551 ability among Bacillus spp. outside the Bacillus cereus group. Appl. Environ.

552 Microbiol. 71, 1178–1183 (2005).

553 37. Serrano, L., Manker, D., Brandi, F. & Cali, T. The Use of Bacillus subtilis QST 713

554 and Bacillus pumilus QST 2808 as Protectant Fungicides in Conventional Application

555 Programs for Black Leaf Streak Control. Acta Hortic. 986, 149–156 (2013).

556 38. Tactacan, G. B., Schmidt, J. K., Miille, M. J. & Jimenez, D. R. A Bacillus subtilis

557 (QST 713) spore-based probiotic for necrotic enteritis control in broiler chickens. J.

558 Appl. Poult. Res. 22, 825–831 (2013).

559 39. Ruiu, L. Brevibacillus laterosporus, a Pathogen of Invertebrates and a Broad-Spectrum

560 Antimicrobial Species. Insects 4, 476–492 (2013).

561 40. Osimani, A. et al. Occurrence of transferable antibiotic resistances in commercialized

562 ready-to-eat mealworms (Tenebrio molitor L.). Int. J. Food Microbiol. 263, 38–46

563 (2017).

564 41. Madigan, M. T., Martinko, J. M., Dunlap, P. V & Clark, D. P. Brock Biology of

565 microorganisms. (Pearson/Benjamin Cummings, 2009).

566 42. Ceuppens, S., Boon, N. & Uyttendaele, M. Diversity of Bacillus cereus group strains is

567 reflected in their broad range of pathogenicity and diverse ecological lifestyles. FEMS

568 Microbiology Ecology 84, 433–450 (2013).

569 43. Zhuang, K. et al. Typing and evaluating heat resistance of Bacillus cereus sensu stricto

570 isolated from the processing environment of powdered infant formula. J. Dairy Sci.

571 102, 7781–7793 (2019).

572 44. Guinebretière, M. H. et al. Ecological diversification in the Bacillus cereus Group.

573 Environ. Microbiol. 10, 851–865 (2008).

574 45. Fricker, M., Messelhäußer, U., Busch, U., Scherer, S. & Ehling-Schulz, M. Diagnostic

575 real-time PCR assays for the detection of emetic Bacillus cereus strains in foods and

576 recent food-borne outbreaks. Appl. Environ. Microbiol. 73, 1892–1898 (2007).

577 46. Ehling-Schulz, M., Frenzel, E. & Gohar, M. Food-bacteria interplay: Pathometabolism

578 of emetic Bacillus cereus. Front. Microbiol. 6, 704 (2015).

579 47. Hoton, F. M. et al. Family portrait of Bacillus cereus and Bacillus weihenstephanensis

580 cereulide-producing strains. Environ. Microbiol. Rep. 1, 177–183 (2009).

581 48. Agata, N., Ohta, M. & Yokoyama, K. Production of Bacillus cereus emetic toxin

582 (cereulide) in various foods. Int. J. Food Microbiol. 73, 23–27 (2002).

583 49. Finlay, W. J. J., Logan, N. A. & Sutherland, A. D. Bacillus cereus produces most

584 emetic toxin at lower temperatures. Lett. Appl. Microbiol. 31, 385–389 (2000).

585 50. Rouzeau-Szynalski, K., Stollewerk, K., Messelhäusser, U. & Ehling-Schulz, M. Why

586 be serious about emetic Bacillus cereus: Cereulide production and industrial

587 challenges. Food Microbiol. 85, 103279 (2020).

588 51. Diez-Valcarce, M., Kovač, K., Cook, N., Rodríguez-Lázaro, D. & Hernández, M.

589 Construction and Analytical Application of Internal Amplification Controls (IAC) for

590 Detection of Food Supply Chain-Relevant Viruses by Real-Time PCR-Based Assays.

591 Food Anal. Methods 4, 437–445 (2011).

592 52. Diez-Valcarce, M. et al. Occurrence of Human Enteric Viruses in Commercial Mussels

593 at Retail Level in Three European Countries. Food Environ. Virol. 4, 73–80 (2012).

594 53. Di Bartolo, I., Angeloni, G., Ponterio, E., Ostanello, F. & Ruggeri, F. M. Detection of

595 hepatitis E virus in pork liver sausages. Int. J. Food Microbiol. 193, 29–33 (2015).

596 54. Dijk, R. et al. Microbiologie van voedingsmiddelen: Methoden, principes en criteria.

597 (MYbusinessmedia, 2015).

598 55. Lievens, B. et al. Design and development of a DNA array for rapid detection and

599 identification of multiple tomato vascular wilt pathogens. FEMS Microbiol. Lett. 223,

600 113–122 (2003).

601 56. Yoon, S. H. et al. Introducing EzBioCloud: A taxonomically united database of 16S

602 rRNA gene sequences and whole-genome assemblies. Int. J. Syst. Evol. Microbiol. 67,

603 1613–1617 (2017).

604 57. Ueda, S., Yamaguchi, M., Iwase, M. & Kuwabara, Y. Detection of Emetic Bacillus

605 cereus by Real-Time PCR in Foods. Biocontrol Sci. 18, 227–232 (2013).

606 Figures

607 608 609 Figure 1 Mean Diamonds plots representing the microbial counts in log cfu/g of the samples 610 investigated in experiment 1. Figure 1A and 1B display the total viable aerobic and anaerobic 611 counts, respectively, Figure 1C and 1D display aerobic and anaerobic bacterial endospore 612 counts, respectively. ● and ▲ represent the three replicates of yellow mealworm (YM) and 613 house cricket (HC) samples, respectively. Sample names consist of the insect species followed 614 by the rearing company number and the batch number. Anaerobic bacterial endospore counts 615 for YM 1.1 were not determined. The horizontal black line represents the mean of all samples. 616 The top and bottom of each diamond represent the 100% confidence interval for each sample 617 and the line across the middle of each diamond represents the sample mean. 95% confidence 618 marks appear as lines above and below the sample mean. 619 a,b,cMean counts from different samples with the same superscript are not statistically different 620 (p > 0.05). $Only two replicates were displayed.

621

622

623 Figure 2 (2A) Number of bacterial endospore isolates from yellow mealworms and house 624 crickets corresponding to the bacterial genera identified by 16 S rRNA gene sequencing and 625 (2B) presence of the cesA gene, in B. cereus group isolates, as determined by qPCR.

626 Tables

Table627 1 Primers and probes employed in this study. Target Oligonucleotide (name) Sequence (5’ – 3’) Forward primer (27F) AGA GTT TGA TCC TGG CTC AG 16 S rRNA gene Reverse primer (1492R) TAC GGY TAC CTT GTT ACG ACT T

B. cereus genes Forward primer (panCF) TYG GTT TTG TYC CAA CRA TGG panC43 Reverse primer (panCR) CAT AAT CTA CAG TGC CTT TCG

Forward primer (cesAF) CAC GCC GAA AGT GAT TAT ACC AA cesA45 Reverse primer (cesAR) CAC GAT AAA ACC ACT GAG ATA GTG

Viruses Forward primer (HAV68) TCA CCG CCG TTT GCC TAG HAV52 Reverse primer (HAV240) GGA GAG CCC TGG AAG AAA G Probe (HAV150) 6FAM-CCT GAA CCT GCA GGA ATT AA-MGBNFQ*

Forward primer (JVHEVF) GGT GGT TTC TGG GGT GAC HEV52 Reverse primer (JVHEVR) AGG GGT TGG TTG GAT GAA Probe (JVHEVP) 6FAM-TGA TTC TCA GCC CTT CGC-BHQ*

Forward primer (QNIF2d) ATG TTC AGR TGG ATG AGR TTC TCW GA NoV GII52 Reverse primer (COG2R) TCG ACG CCA TCT TCA TTC ACA Probe (QNIFS) 6FAM-AGC ACG TGG GAG GGC GAT CG-BHQ*

IAC51 IAC probe (PrfAP) VIC-CCA TAC ACA TAG GTC AGG-MGBNFQ* *6FAM:628 6-carboxyfluorescein, MGB-NFQ: minor groove binder non-fluorescent quencher, BHQ: black629 hole quencher, VIC: 2’-chloro-7’-phenyl-1,4-dichloro-6-carboxyfluorescein. 630

b c A b B b,c

a,b b,c a,b,c a,b

a a,b a a,b a,b a a a

b C c D b b,c b a,b a,b,c a,b,c a,b,c a,b a,b a,b a,b a a bioRxiv preprint doi: https://doi.org/10.1101/2020.04.22.055236; this version posted April 23, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.